16:04:28 #startmeeting networking_ml2 16:04:29 Meeting started Wed Mar 1 16:04:28 2017 UTC and is due to finish in 60 minutes. The chair is rkukura. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:04:30 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:04:32 The meeting name has been set to 'networking_ml2' 16:04:40 Hi everyone! 16:04:45 Hi 16:05:05 #topic Agenda 16:05:21 #link https://wiki.openstack.org/wiki/Meetings/ML2#Meeting_March_1.2C_2017 16:05:48 anyone want to add/remove anything? 16:06:10 #topic Announcements 16:06:21 I don’t have any announcements - anyone else? 16:06:58 #topic PTG post mortem 16:07:20 Did everyone enjoy the PTG last week in Atlanta? 16:07:27 yes :) 16:07:38 Great, so did I 16:08:18 I think we had some productive ML2-related discussions 16:08:30 me too 16:08:33 it was fun 16:08:55 yes - I think ML2 sessions were great 16:09:21 it is alway good to meet f2f 16:09:44 yeah the discussions were very helpful 16:09:56 ML2-related topics I recall being discussed include the live migration work, enforcing extension semantics, SR-IOV and gluon 16:10:02 I’m sure I’m missing some 16:10:58 Lets at least summarize next steps on these (and any I missed) 16:11:10 we’ll cover live migration next on the agenda 16:11:41 rkukura : I think your list is accurate 16:11:59 I plan to draft a RFE for extension semantic enforcement this week - hopefully we’ll be able to discuss it here next week 16:12:28 that would be fantastic 16:13:22 For live migration I am almost done with the apis 16:13:24 I’m not sure what all was discussed regarding SR-IOV - can anyone summarize that? 16:13:41 next step is to do the integration with the ml2 16:14:04 dasanind: thanks 16:15:15 Anyone know next steps regaring SR-IOV? Anything beyond adding/enhancing mechanism drivers? 16:15:28 rkukura : regarding SR-IOV, there is an RFE for extensions for SR-IOV driver 16:16:14 Sukhdev_: thanks - lets track that in next weeks agenda - do you have a link handy? 16:16:33 * Sukhdev_ looking 16:16:36 I recall some discussion of QinQ - was this for SR-IOV? 16:19:14 we’ll include the link in next week’s agenda 16:19:38 #link - https://bugs.launchpad.net/neutron/+bug/1662650 16:19:38 Launchpad bug 1662650 in neutron "[RFE] Advance configuration of SR-IOV ports- api extension" [Wishlist,New] 16:20:08 Sukhdev_: thanks! 16:20:09 yes QinQ discussion was for SR-IOV related 16:20:31 Sukhdev_: Is it covered by this RFE, or separate? 16:20:50 I do not believe that is covered by RFE 16:21:02 I do see mentions of “outer vlan tag” 16:21:34 rkukura: what is QinQ? 16:21:37 but I think there was a separate discussion 16:21:47 dasanind: Its vlans-within-vlans 16:22:08 I thought that was to be addressed by type driver 16:22:19 rkukura: thanks 16:22:24 I mean new type driver to deal with QinQ 16:22:53 Sukhdev_: right, the discussion I caught the end of seemed to be about allocating tags for QinQ, which might require type driver support 16:23:37 correct 16:23:39 OK, lets see if we can find an RFE for this, and include it in next week’s agenda as well 16:24:41 So, for Gluon, my understanding is that there will be an (or is) an RFE for a “network type” attribute that would select which of multiple core plugins would handle networks of each type 16:25:37 So ML2 could handle traditional networks, and other core plugins could handle specialized networks 16:26:14 Sukhdev: looks like Sukhdev_ left - are you still here? 16:27:16 rkukura: I missed this discussion so I do not have any information about this :( 16:27:33 We might have lost Sukhdev, who I was hoping could provide some details on this 16:28:37 OK, lets cover that next week as well 16:28:46 ok 16:28:48 Anything else on the PTG last week? 16:29:33 sorry - did not realize I was disconnected 16:30:10 Sukhdev_: can you summarize where things stand on RFEs related to Gluon? 16:30:41 Actually I was typing those and did not realize I was disconnected :-):-) 16:31:29 I filed an RFE for network type - it is out there and we used it for the discussion at PTG 16:31:59 OK, lets link this into next week’s agenda as well 16:32:03 and there is a development effort going on for the plugin - 16:32:13 Gluon plugin in being build on top of ML2 plugin - 16:32:53 My understanding is that this network type would be used to dispatch to one of several core plugins 16:33:12 yes - your understanding is correct rkukura 16:33:43 So would gluon re-use certain aspects of ML2 while ML2 itself would handle traditional networks? 16:34:10 the initial use case is to have a service chain when more than one SDN controllers are involved - each orchestrating different service 16:34:27 yes - that is correct 16:35:07 the idea is to be able to hybrid deployments - 16:35:52 traditional networks, PnP networks and service chained networks - all orchestrated by multiple SDN controllers, along with Neutron 16:36:07 in other words - mother of all :-):-) 16:36:26 challenge is to coordinate everything 16:36:53 right - there is work flow being designed as well - 16:37:05 all driven by an auto-generated API 16:37:17 generated from what? 16:37:27 the thing is very involved - 16:37:43 If I tell you all, then I will have to kill you :-):-) 16:37:57 just kidding :-) 16:38:01 :) 16:38:26 there is a proton service plugin is being designed which works with Yang models - 16:38:44 these yang models auto-generate APIs - 16:38:55 which are then fed to Gluon Plugin 16:39:12 some of these APIs are for new type of network 16:39:27 and some of the APIs will be traditional neutron APIs 16:39:51 hence, the Gluon Plugin needs to deal with both ML2 and new network type use cases 16:40:32 Sukhdev_: I see https://wiki.openstack.org/wiki/Gluon - is this now an official OpenStack project, big tent, stadium, or what? 16:40:45 We only discussed the Gluon Plugin part at PTG 16:41:17 most of the other part is all new and will be covered under a new service plugin 16:42:10 looks like we lost Sukhdev/Sukhdev_ again 16:43:22 OK, we should move on with the agenda 16:43:31 anything else on the PTG? 16:44:25 he’s back 16:44:50 ok, I’d like to wrap up this part of the agenda - anything else on Gluon and/or the PTG for today? 16:45:22 nothing else from my side 16:45:25 ok 16:45:35 nothing from me - 16:45:48 I might get disconnected again - 16:46:00 Sukhdev_: Gluon is a separate plugin or is it in neutron tree 16:46:28 dasanind : separate plugin in its own repo 16:46:46 dasanind : once all done, will be brought under neutron tent 16:46:52 stadium 16:46:55 Sukhdev_: Is gluon official big tent project, part of neutron stadium, or what? 16:47:06 not yet 16:47:15 not official yet? 16:47:18 all experimental stuff at this point 16:47:23 most projects are 16:47:33 is there a RFE for this? 16:48:12 dasanind : no RFE for Gluon 16:48:30 ok 16:48:32 We already touched on the live migration port binding support, and on RFEs, so we can move to open discussion and wrap up 16:48:32 it will be done in multiple pieces 16:48:40 #topic Open Discussion 16:49:32 Need reviews on the api's for live migration https://review.openstack.org/#/c/414251 16:50:19 dasanind: Thanks for the reminder 16:50:59 anything else? 16:51:12 I reviewed the older version - will look at the updated one 16:51:16 nothing from me 16:51:17 nothing else from me 16:51:19 ok, see you here next week 16:51:24 thank you rkukura and Sukhdev_ 16:51:28 thanks rkukura 16:51:31 thanks 16:51:32 #endmeeting